AI at the Zoo: Introduction

I am kicking off a new AI at the Zoo Series! Those of you who have known me for a while know that I love animals (especially pandas) and also making silly apps with no purpose. Recently, I have been learning more about AI and the Power Platform AI Builder. During this exploration I found the improved Image Classification powered by Lobe.

Lobe is a cool tool for training Object Classification models. You download it on your machine and then begin training. After the initial training, it will continue to update and re-train as you add more images and test your model. Then you can export this to the Power Platform to use in Power Apps or Power Automate.

So, my goal of this app was to:

1. Allow the user to take a picture of an animal
2. Determine which animal this was. I trained it to recognize pandas, zebras, elephants, lions and tigers. These animals were selected as some live at the Washington, DC Smithsonian National Zoo.
3. Save the label and picture to Dataverse for future reference

Over the next few days, I will walk through setting up and exporting the model, designing the Power App and testing in the real world!

If you want to follow along, you should do the following things to prepare:

1. Download Lobe
2. Set up a Trial Environment. I recommend following the steps in this oldie-but-a-goodie post from Nick Doelman. This will walk you through creating a new browser profile, setting up an E3 Office trial, adding a Sales Trial.
3. Enable the AI Builder trial on the Sales Trial environment. This can be done in the Maker Portal (Make.PowerApps.com), AI Builder area on the left-hand side.

You can also check out the Power Tips video where we talked about Lobe and AI Builder for more of an introduction.

Spoiler Alert: If you want to skip to the good part, check out my #AiAtTheZoo tweets from my live testing.

Watch for more posts in the AI at the Zoo series or check out Events where you can find me presenting on AI or other Power Platform topics!